Abstract
Objectives: Staphylococcus aureus bacteraemia (SAB) is a common infection associated with significant short-term mortality. Little is known about long-term prognosis. The aim of this study was to determine one-year all-cause mortality and infection-related mortality and associated predictors.
Methods: Data from 303 consecutive patients with SAB were prospectively collected from March 2011 to February 2014. All patients were followed one year or until death.
Results: One-year all-cause- and infection-related mortality were 36.7% and 20.8%, respectively. For all-cause mortality, in multivariable logistic regression analysis, age 70–79 years (OR 3.9; 95% CI 1.7–9.1; p = .001), Charlson Comorbidity index ≥3 (OR 6.9; 95% CI 2.7–17.3; p < .001), healthcare-associated infection (OR 2.3; 95% CI 1.1–4.9; p = .03) and severe sepsis (OR 3.6; 95% CI 1.8–7.1; p < .001) were independent predictors of outcome. For infection-related mortality, the predictors were similar, except for healthcare-associated infection that lost significance. The vast majority (89%) of infection-related deaths occurred within 30 days.
Conclusions: This study demonstrates additional significant all-cause mortality in patients with SAB beyond 30 days to one year, mainly driven by high age and comorbidity. As a result, SAB can be considered an indirect marker of high risk of death in these patients. Follow-up beyond 30 days does not add significant information with respect to infection-related mortality.
Introduction
Staphylococcal aureus bacteraemia (SAB) is one of the most common bloodstream infections with annual incidence rates of 20–30/100 000 population [Citation1–3]. It represents a myriad of different clinical presentations and disease courses influencing treatment and outcome. During the last decades the prognosis has steadily improved to a 30-day all-cause mortality rate of approximately 20–25% [Citation3–5]. Multiple factors independently affecting short-term mortality have been identified in several studies including age, comorbidity, mode of acquisition, presence of metastatic foci, methicillin resistance, severity of sepsis and delayed appropriate antibiotic therapy [Citation6]. Given the heterogeneous nature of SAB infections with many patients having long and complicated disease courses, 30-day outcome may underestimate mortality, hence knowledge of long-term prognosis is important. There is a paucity of studies evaluating one-year outcome, mostly retrospective and with mortality rates ranging from 30% to 62% [Citation7–12]. In a recent review of ten studies reporting long-term prognosis, the median one-year mortality was 46% [Citation13].
In addition, only a few studies have reported infection-related mortality. The direct contribution of the infection to outcome is important to get a more complete understanding of factors that determine both short- and long-term prognosis.
The principle aim of the present study was to assess one-year all-cause mortality and potential predictors in a prospective cohort study of patients with SAB. Secondary aims were to describe infection-related mortality, determine potential predictors of outcome and to compare them with all-cause mortality.
Methods
Study population
The cohort has been described in details elsewhere [Citation3]. In brief, all patients ≥18 years of age with a first episode of SAB admitted to Akershus University Hospital from March 2011 to February 2014 were prospectively included in the study. Polymicrobial infections were excluded. Baseline characteristics and clinical data were obtained from medical records and bedside clinical examination. Mortality data from the Civil Registry are automatically updated in the electronic medical charts.
Definitions
We defined infection-related mortality as death that could be attributed to SAB, either as the immediate or indirect cause no matter the severity and prognosis of concurrent comorbid conditions. In practice, to be counted as infection-related mortality, patients had ongoing unequivocal clinical and/or biochemical signs of infection at the time of their death. All cases were reviewed by the two infectious disease specialists (Arne Nørgaard Eskesen, Margrethe Astrup Belle). When the cause of death was not straightforward it was discussed among them and a decision made thereafter.
Hospital-acquired infection was defined by a positive S. aureus blood culture obtained at least 48 hours after admission. Healthcare-associated and community-acquired infection was defined by a positive blood culture obtained within 48 h of admission. Patients with healthcare-associated infection had received significant medical treatment within 90 days prior to hospitalization [Citation14].
Severe sepsis was defined as sepsis with organ dysfunction and septic shock as sepsis with circulatory collapse [Citation15]. Charlson Comorbidity Index was used to quantify comorbidity [Citation16]. Reinfection was defined as a new episode of SAB after a previous clinical og biochemical cured infection whereas relapse was defined as recurrence of an insufficiently treated infection irrespective of any time interval.
Statistical analysis
All statistical analyses were conducted using the Statistical Package for the Social Sciences (SPSS) version 23 (SPSS Inc., Chicago, IL, USA). Patients were followed one year or until death.
Predictors of mortality were assessed by univariate and multivariable logistic regression and presented with odds ratios, 95% confidence intervals and p values.
The multivariable models were built sequential with variables entered one at a time and the significance and the variance contributing to the model assessed at each step. A significance level of 0.05 was used to remove variables from the model. Age, gender, Charlson comorbidity index, mode of acquisition, severity of sepsis and timing of antibiotics were considered as potential explanatory variables.
Primary and secondary foci of infection were not included in the model due to well-known strong correlation with mode of acquisition and thus high likelihood of collinearity.
Results
During the study period 353 adult patients had blood cultures positive for S. aureus.
Fifty patients were excluded, due to polymicrobial infection (n = 29), not living in the catchment area (n = 7), contamination (n = 5) or not being available for inclusion (n = 9).
The final study cohort consisted of 303 patients with a mean age of 71 years (range 18–95 years) and 61% were men. Five patients (1.8%) showed methicillin-resistant (MRSA) infection. Relapse and reinfection occurred in 18 patients (5.9%) and 12 patients (4.0%), respectively.
The infection was hospital-acquired in 84 patients (28%), healthcare-associated in 130 patients (43%) and community-acquired in 89 patients (29%) (). Two hundred and thirty-four patients (77%) had one or more comorbid conditions.
Table 1. Patient and clinical characteristics of 303 patients with Staphylococcus aureus bacteraemia.
The most common primary focus of infection was intravascular catheter, 73 patients (24%), skin and soft tissue, 67 patients (22%), or unknown focus, 96 patients (32%). Osteomyelitis and/or septic arthritis was the most prevalent secondary focus prevalent in 81 patients (27%). Sixty-six patients (22%) had severe sepsis or septic shock. One hundred and seventy seven patients (42%) received antibiotics within 48 hours of debut of symptoms.
Overall 63 patients (20.8%) died within 30 days (). Of those, 56 deaths (89%) were infection-related. One hundred and ten patients (36.7%), died within one-year.
Table 2. All-cause mortality and infection-related mortality in 303 patients with Staphylococcus aureus bacteraemia.
Factors associated with one-year all-cause mortality
In univariate analysis age ≥70 years, Charlson comorbidity index ≥1, hospital-acquired infection, healthcare-associated infection and severe sepsis were associated with mortality (). In multivariate analysis age ≥70 years, Charlson comorbidity index ≥3, healthcare-associated infection and severe sepsis remained independent predictors. Highest impact on mortality was observed for age ≥80 years (OR 6.5, 95% CI 2.9–14.6; p < .001) and Charlson comorbidity index ≥3 (OR 6.9; 95% CI 2.7–17.3, p < .001). Treatment with appropriate antibiotics within 48 h of debut of symptoms did not affect outcome.
Table 3. Predictors of 1-year all-cause mortality in 303 patients with Staphylococcus aureus bacteraemia.
Restricting the analysis to the 240 patients who were alive at day 30, age ≥80 years and Charlson comorbidity index ≥3 remained significant (OR 3.0, 95% CI 1.1–8.0; p = .03 and OR 6.8, 95% CI 2.1–21.7; p = .001) respectively) whereas healthcare-associated infection and severe sepsis lost predictive value.
Factors associated with one-year infection-related mortality
Sixty-three patients (20.8%) died within one-year related to their initial S. aureus infection. The vast majority, 56 patients (89%), died within the first 30 days and only seven patients died between day 30 and one year.
Forty-seven patients who died due to other causes than infection were excluded from the analysis. In the univariate analysis age ≥70 years, Charlson comorbidity index ≥3, healthcare-associated infection and severe sepsis were associated with mortality and all remained independent predictors in the multivariable model except for healthcare-associated infection (). Age ≥80 years (OR 19.3, 95% CI 5.7–65.5; p < .001) and severe sepsis (OR 8.8, 95% CI 3.9–19.8; p < .001) had highest impact on mortality.
Table 4. Predictors of one-year infection-related mortality in 256 patients with Staphylococcal aureus bacteraemia.
Discussion
In this prospective study of 303 patients with SAB we demonstrated a significant additional all-cause mortality beyond 30 days steadily increasing to 36.7% at one year. This is lower than the median 46% found in a recent review of 10 studies, but different study designs and time periods make more detailed comparisons difficult [Citation13].
Numerous studies have examined short-term predictors of mortality in patients with SAB. Age, comorbidity and severity of infection have consistently been associated with an increased risk of death whereas gender, mode of acquisition and timing of antibiotics have led to conflicting results [Citation6]. Our findings demonstrates that age ≥70 years and significant comorbidity with Charlson comorbidity index ≥3 are the strongest predictors of one-year all-cause mortality. When patients that died within 30 days of the diagnosis of SAB were excluded from the analysis severe sepsis did not predict long-term prognosis, reflecting that severe infection largely influences short-term outcome.
Mode of acquisition in the form of healthcare-associated infection are likely to be a tenuous predictor and lost its significance after 30 days, in line with several studies of short-term prognosis that failed to demonstrate an association. The rather broad definition of healthcare-associated infection combined with the substantial heterogenenicity of systemic S. aureus infections can possibly explain the weakness of the association.
Two previous studies of long-term prognosis found female gender to be an independent predictor of mortality with hazard ratios of 1.3 and 1.08, respectively [Citation10,Citation11]. Potential explanations are exclusively speculative. We could not confirm this association in our study and the effect size, if any, are probably minimal [Citation6].
Prompt treatment with appropriate antibiotics represents a cornerstone in the treatment of patients with sepsis and delay of insertion is associated with increasing mortality. In patients with SAB however, findings are less clear and conflicting results have been reported [Citation17–20]. Two studies of long-term prognosis failed to find a significant association using 48 hours after the index culture was drawn as cut off between appropriately timed and delayed treatment with antibiotics [Citation7,Citation10].
Similarly, the duration of symptoms until treatment with antibiotics to which the pathogen was sensitive, did not influence outcome in our study when using 48 h as cut off. The reasons for these apparently contradictory findings are not clear. Finally, one-year relapse rate was low (5.9%) and thus unlikely to have a significant impact on mortality.
In general, studies of patients with sepsis have consistently shown high crude case fatality rates. A large fraction of these patients are old, fragile and with numerous comorbid conditions, but little attention have been given to estimate infection-related and attributable mortality. This is important, not only to get a more comprehensive knowledge of these infections, but especially to be able to gauge the impact of interventions such as treatment with antibiotics [Citation21]. Lack of widely accepted definitions and methodological difficulties probably contribute to the scarcity of data. We defined infection-related mortality as death that could be assigned to SAB, either as the immediate or indirect cause, reflecting the clinicians judgement of what led to death, unadjusted for concurrent comorbid conditions. This perspective adds valuable information in studies of long term prognosis too, where mortality due to SAB seems to sharply decline after one month.
In fact, the main drivers of death after 30 days are older age and comorbidity and although not causally related, SAB can be considered as a marker of high risk of death in these patients. Younger age and having no comorbidity carries a much lower risk of mortality. The youngest patient without comorbidity that died within one year of SAB was 73 years old in our cohort.
As infection-related death due to SAB after 30 days was rare, predictors of infection-related short-term and one-year mortality are similar. Thus, for future studies follow-up beyond 30 days are unlikely to add significant additional information with respect to infection-related mortality.
Attributable mortality, on the other hand, has been defined as the direct contribution to death from infections after adjusting for underlying illnesses [Citation21]. It offers an additional perspective to infection-related mortality, but requires a precisely matched control group. In a nested case-cohort study in-hospital attributable mortality for hospital-acquired SAB has been estimated to 15% [Citation22].
The strength of our study is the prospective design ensuring the quality of the clinical data collected during a three-year period by two infectious diseases specialists. In addition, more than 97% of eligible patients with SAB were included and automatic updating of the electronic medical charts with mortality data from the civil registry make certain that the case fatality rates are precise.
Determining if the cause of death is infection-related or not, can in some cases be difficult and therefore to a certain extent subjective and prone to bias and that is a limitation of the study. The vast majority of cases were straightforward though and we believe that the number of potential misclassifications is low. In addition, after discussion of equivocal cases there were no disagreement between the two infectious disease specialists.
In conclusion, this study demonstrates that after day 30, the serious one-year prognosis in patients with SAB is mainly driven by high age and significant comorbidity. In line with this, the vast majority (89%) of the infection-related mortality occurs within the first 30 days.
Ethical approval
The study was conducted in accordance with good clinical practice and the Declaration of Helsinki, and approved by the regional ethics committee of The South-Eastern Regional Health Authority (2009/2149).
Informed consent
Informed consent was obtained from all individual participants included in the study.
| Abbreviations | ||
| SAB | = | Staphylococcus aureus bacteraemia |
| CCI | = | Charlson comorbidity index |
Disclosure statement
No potential conflict of interest was reported by the authors.
References
- Mejer N, Westh H, Schønheyder HC, et al. Stable incidence and continued improvement in short term mortality of Staphylococcus aureus bacteraemia between 1995 and 2008. BMC Infect Dis. 2012;12:260.
- Laupland KB, Lytikäinen O, Søgard M, et al. The changing epidemiology of Staphylococcus aureus bloodstream infection: a multinational population-based surveillance study. Clin Microbiol Infect. 2013;19:465–471.
- Blomfeldt A, Eskesen AN, Aamodt HV, et al. Population-based epidemiology of Staphylococcus aureus blodstream infection: clonal complex 30 genotype is associated with mortality. Eur J Clin Microbiol Infect Dis. 2016;35:803–813.
- Jacobsson G, Gustafsson E, Andersson R. Outcome for invasive Staphylococcus aureus infections. Eur J Clin Microbiol Infect Dis. 2008;27:839–848.
- Turnidge JD, Kotsanas D, Munckhof W, et al. Staphylococcus aureus bacteraemia: a major cause of mortality in Australia and New Zealand. Med J Aust. 2009;191:368–373.
- van Hal SJ, Jensen SO, Vaska VL, et al. Predictors of mortality in Staphylococcus aureus bacteremia. Clin Microbiol Rev. 2012;25:362–386.
- Fätkenheuer G, Preuss M, Salzberger B, et al. Long-term outcome and quality of care of patients with Staphylococcus aureus bacteremia. Eur J Clin Microbiol Infect Dis. 2004;23:157–162.
- Hanses F, Spaeth C, Ehrenstein BP, et al. Risk factors associated with long-term prognosis of patients with Staphylococcus aureus bacteremia. Infection. 2010;38:465–470.
- Jacobsson G, Nasic S. Long-term outcome of invasive Staphylococcus aureus infections. Scand J Inf Dis. 2012;44:350–354.
- Yahav D, Yassin S, Shaked H, et al. Risk factors for long-term mortality of Staphylococcus aureus bacteremia. Eur J Clin Microbiol Infect Dis. 2016;35:785–790.
- Gotland N, Uhre ML, Mejer N, et al. Long-term mortality and causes of death associated with Staphylococcus aureus bacteremia. A matched cohort study. J Infect. 2016;73:346–357.
- Huggan PJ, Wells JE, Browne M, et al. Population-based epidemiology of Staphylococcus aureus bloodstream infection in Canterbury, New Zealand. Intern Med J. 2010;40:117–125.
- Asgeirsson H, Thalme A, Weiland O. Staphylococcus aureus bacteraemia and endocarditis – epidemiology and outcome: a review. Infect Dis (Lond). 2018;50:175–192.
- Friedman ND, Kaye KS, Stout JE, et al. Healthcare-associated bloodstream infections in adult: a reason to change the accepted definition of community-acquired infections. Ann Intern Med. 2002;137:791–797.
- Levy MM, Fink MP, Marshall JC, et al. 2001 SCCM/ESICM/ACCP/ATS/SIS International Sepsis Definitions Conference. Intensive Care Med. 2003;29:530–538.
- Charlson ME, Pompei P, Ales KL, et al. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987;40:373–383.
- Lodise TP, McKinnon PS, Swiderski L, et al. Outcomes analysis of delayed antibiotic treatment for hospital-acquired Staphylococcus aureus bacteremia. Clin Inf Dis. 2003;36:1418–1423.
- Paul M, Kariv G, Goldberg E, et al. Importance of appropriate empirical antibiotic therapy for methicillin-resistant Staphylococcus aureus bacteraemia. J Antimicrob Chemother. 2010;65:2658–2665.
- Kim SH, Park WB, Lee CS, et al. Outcome of inappropriate empirical antibiotic therapy in patients with Staphylococcus aureus bacteraemia: analytical strategy using propensity scores. Clin Microbiol Infect. 2006;12:13–21.
- Ammerlaan H, Seifert H, Harbarth S, et al. Adequacy of antimicrobial treatment and outcome of Staphylococcus aureus bacteremia in 9 Western European countries. Clin Infect Dis. 2009;49:997–1005.
- Wenzel RP, Gennings C. Residual attributable mortality, a new concept for understanding the value of antibiotics in treating life-threatening acute infections. Antimicrob Agents Chemother. 2010;54:4956–4960.
- Brady M, Oza A, Cunney R, et al. Attributable mortality of hospital-acquired bloodstream infections in Ireland. J Hosp Infect. 2017;96:35–41.